You are on page 1of 10

C. R.

Physique 10 (2009) 197206


Laser acceleration of particles in plasmas / Acclration laser de particules dans les plasmas
Ion acceleration with high-intensity lasers and application
to isochoric heating
Erik Lefebvre

, Michal Carri, Rachel Nuter


CEA, DAM, DIF, Bruyres-le-Chtel, 91297 Arpajon cedex, France
Available online 28 April 2009
Abstract
Very nonlinear mechanisms are at play in the acceleration of ions with high-intensity laser pulses, and the properties of the
resulting particle beams depend on many parameters. We rst explore, in this article, the dependence on the target density prole
and on the laser pulse duration. We next show numerically that using targets on which thin, low-density proton layers are deposited,
it should be possible to produce proton beams suitable for heating a solid aluminum target to the GJ/kg level over a depth of several
micrometers. To cite this article: E. Lefebvre et al., C. R. Physique 10 (2009).
2009 Acadmie des sciences. Published by Elsevier Masson SAS. All rights reserved.
Rsum
Acclration ionique par laser ultra-intense et application au chauffage isochore. Lacclration dions laide dimpulsions
laser ultra-intenses met en jeu une physique complexe, et les proprits des faisceaux de particules produits dpendent de nombreux
paramtres. Nous explorons tout dabord dans cet article les dpendances avec le prol de densit de la cible et la dure de
limpulsion laser. Dans un second temps, nous montrons laide de simulations numriques quil devrait tre possible, en utilisant
des cibles sur lesquelles sont dposes des couches minces et peu denses de protons, dacclrer un faisceau dions susceptible de
chauffer une cible daluminium jusqu 1 GJ/kg sur une profondeur de plusieurs micromtres. Pour citer cet article : E. Lefebvre
et al., C. R. Physique 10 (2009).
2009 Acadmie des sciences. Published by Elsevier Masson SAS. All rights reserved.
Keywords: Ion acceleration; Ultrahigh-intensity lasers; Isochoric heating
Mots-cls : Acclration ionique ; Laser ultra-intense ; Chauffage isochore
1. Introduction
In the last decades, improvements in the laser Chirped Pulse Amplication technology have allowed the construc-
tion of powerful laser facilities able to deliver on target focused intensities up to several 10
21
W/cm
2
, or even several
10
22
W/cm
2
[1]. These intense laser pulses can be used to efciently accelerate high quality proton or ion beams off
the rear surface of thin, solid targets. These particle beams are laminar [2], collimated [3] and with m virtual source
*
Corresponding author.
E-mail address: erik.lefebvre@cea.fr (E. Lefebvre).
1631-0705/$ see front matter 2009 Acadmie des sciences. Published by Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.crhy.2009.03.005
198 E. Lefebvre et al. / C. R. Physique 10 (2009) 197206
size, and the maximum particle energy can reach several tens of MeV per nucleon. They can be used in a wide range
of applications such as imaging in plasma physics [4], the production of warm dense matter [5], or in the medical eld
with proton therapy [6].
Compared to heating with intense electron beams, protons could offer an interesting alternative for the production
of warm dense matter, or even to reach ignition condition in compressed thermonuclear fuel [7]. Indeed, for a similar
kinetic energy ux, a proton beam transports less current with heavier particles, and is therefore less affected by
nonlinear, self-generated electromagnetic elds. For a given incident particle kinetic energy, the stopping power of
ions is also larger than that of electrons, enabling more localized deposition and heating. Finally, some indications
of control have been demonstrated with proton and carbon beams, such as focusing [5] and energy selection [810].
Alternative schemes to produce monoenergetic ion distributions have been proposed, based on numerical simulations
and theory [11,12], but still await experimental conrmation. All these methods could be used for optimal control of
energy deposition in a secondary target. Yet even though proton acceleration with intense lasers has been the subject of
much interest recently, only a limited number of experimental results have been published regarding proton-induced
isochoric heating [5,1315].
Optimization of ion acceleration is often considered from the point of view of increased energy or larger coupling
efciency from laser light to ion kinetic energy. Other source properties also have to be optimized with respect to
specic applications. For isotope production, for instance, spectral quality of the proton beam is of lesser importance
than efciency of the laser-to-ion coupling [16]. For application to isochoric heating, two goals can be pursued. For
the rst one, one will look for the highest energy deposition into the material of interest; this can be achieved, for
instance, by focusing the accelerated ion beam [5], or by using low-energy ions that will deposit their energy in a
thinner layer of matter [13]. The other option for optimization might be to look for the largest heated region, and with
the most uniform energy deposition across the heated zone, both along its longitudinal and transverse dimensions, in
order to ease the observation of the heated medium dynamics, and its interpretation. For a nite laser energy, these
two options are likely exclusive.
In this article, we will rst report on some source optimization studies that we performed to try and identify the main
trends of peak ion energy variations with parameters such as target density prole [17] and laser pulse duration [18].
A straightforward path towards uniform energy deposition is obviously to use controlled proton spectra and rely on
the specic properties of proton stopping power through solid matter. In the second half of this article, we will report
on some numerical exploratory studies towards this goal.
2. Dependence of acceleration on target density prole
A question of interest in studies of proton acceleration is the optimization of particle energy for a xed incident
laser energy. Proton acceleration is known to depend on both the mean energy and number of hot electrons that can
be produced by laser absorption in the target. Controlling the plasma density gradient at the laser-irradiated surface
can be one way to control the laser-to-hot electron conversion fraction, as well as the hot electron temperature [19].
This plasma density control can be achieved with an auxiliary, low-intensity heating beam hitting the target prior
to the main, high-intensity pulse. In the following section, we study the possibilities offered by such a two-pulse
conguration.
2.1. Numerical setup
In our numerical experiments, we consider a 500 nm-thick Al target with a variable density gradient on its laser-
irradiated surface, such as could be produced by a low-intensity pulse hitting this target with a variable delay, t ,
before the high-intensity pulse. The interaction of the high-intensity pulse with the Al target was then simulated in
two-dimensional (2D) geometry with the CALDER PIC code [20]. The initial plasma density proles are plotted in
Fig. 1a). They were computed with the 1D Lagrangian hydrodynamics code ESTHER [21] for a low-intensity pulse
of 150 fs duration, focused to 1.1 10
16
W/cm
2
. After 12 ps, the shock wave launched by the low-intensity pulse
breaks out at the rear surface, and likely perturbs proton acceleration [22]. We did not therefore consider delays larger
than 15 ps. Due to computational constraints with the PIC simulations, we have imposed lower and upper bounds on
the density proles, equal to n
c
/175 and 50 n
c
, respectively (n
c
=1.74 10
21
cm
3
is the critical density at the laser
wavelength
0
=0.8 m).
E. Lefebvre et al. / C. R. Physique 10 (2009) 197206 199
Fig. 1. (a) Electron density proles at the target irradiated surface for different time delays. The dot-dashed lines represent the lower and upper
density bounds considered in PIC simulations. (b) Absorption rate versus the time delay for E
in
= 256 mJ (squares) and E
in
= 50 mJ (circles).
t = 0 ps denotes the solid target without preplasma. (c) Maximum proton kinetic energy versus time delay for E
in
= 256 mJ (squares) and
E
in
=50 mJ (circles).
The electron and ion temperatures are initialized to T
e
=1 keV and T
i
=1 eV, respectively. The mesh size is set to
x =y =10.12 nm. To reduce numerical heating down to a low level (<1% of the incident laser energy), 80 parti-
cles per cell are used and the current and charge densities are calculated with quadratic weighting factors. Moreover,
the numerical box (up to 13750 2176 cells) is chosen sufciently large to avoid articial particle absorption at
boundaries: its transverse size (22 m) is larger than three times the focal spot and the longitudinal one (up to 140 m)
is at least equal to 18 times the target thickness. Electrons hitting the box boundaries are reinjected with thermal
momentum as long as ions do not reach the boundaries, in which case both species are absorbed to keep the plasma
globally neutral. The boundary conditions for the electromagnetic elds are periodic along the y axis and open in the
x directions. Finally, a 50 n
c
, 20 nm proton layer is put on the Al target rear surface so as to model the ionized
hydrogen contaminants.
The fast electrons responsible for ion acceleration are generated by an intense, short duration pulse characterized by
two distinct sets of parameters: either a 256 mJ energy, 24 fs duration and a 3.2 m diameter (I
0
=8.610
19
W/cm
2
),
or a 50 mJ energy, 30 fs duration and a 4 m diameter (I
0
=8.6 10
18
W/cm
2
). In both cases, the central wavelength
is 0.8 m and the linearly polarized laser beam is normally incident onto the target, with the laser electric eld (E
y
)
in the simulation plane. These laser parameters are typical of current 10-TW, titaniumsapphire laser systems used in
many laboratories.
2.2. Fast electron generation
We rst focus on the electron acceleration mechanisms occurring during the interaction of the ultra-intense laser
with the target, whether preceded by a plasma or not.
Fig. 1b) plots the absorbed laser energy as a function of the time delay between the low-intensity and relativistic
intensity pulses. Whereas the absorption rate for a steep gradient target (t =0 ps) is identical for both intense pulses
(15%), it is seen to increase with laser energy when a large preplasma (corresponding to t > 5 ps and 10 ps for
E
in
=256 mJ and 50 mJ, respectively) is added in front of the target.
For the shorter delays and steeper plasmas, the two high-intensity pulses show opposing behaviors, with the lower
energy (50 mJ) case showing rst a decreases and then an increase of absorption for t > 2 ps, whereas the high-
energy case exhibits a local maximum around the same delay. This feature suggests that different electron heating
mechanisms are at play in the preplasma. When the normally incident, moderate-intensity laser directly interacts with
the sharp-gradient target, electrons are mostly heated through the relativistic j B mechanism [23]. The Lorentz
force oscillates the electrons inside the target in the longitudinal direction with its evanescent part mainly oscillating
at twice the laser frequency. Electrons are also pulled back towards the plasma surface by the restoring electrostatic
elds. Heating results from the crossings of electron trajectories during this complicated motion [24,25]. As a thin
preplasma is added in front of the solid target, this process becomes less efcient: the absorption rate decreases down
200 E. Lefebvre et al. / C. R. Physique 10 (2009) 197206
to 5.3% for t = 2 ps. Indeed, as the preplasma length increases, the evanescent part of the Lorentz force reaches
regions of lower electron density and can heat a smaller number of particles.
For the high-energy case, on the other hand, the observed local maximum around t =2 ps in the absorption rate is
reminiscent of a theoretical prediction made by Andreev et al. [26]. According to these authors, this optimal preplasma
scale length is that for which the 2
0
-oscillating evanescent part of the Lorentz force is best able to resonantly excite
an electrostatic wave in the vicinity of n
e
= 4n
c
, hence enhancing laser absorption. This point is discussed in more
detail in another publication [17].
Despite those different trends at short delays, both laser energy cases show a steadily increase of absorption with
t for t > 5 ps. In this conguration, the preplasma electrons are accelerated by a standing wave resulting from the
superposition of the incident and reected laser pulses. This most efcient heating mechanism can be related to the
stochastic heating evidenced so far for longer laser durations [27,28]. Sentoku et al. [29] have shown in this case that
the electron kinetic energy and the absorption rate increase with the laser energy and the plasma scale length (l
ss
),
respectively, up to a saturation when l
ss
10
0
. In our conguration, the plasma gradient length is always below this
threshold value, and we observe a monotonous increase in the electron temperature and the absorption rate with the
plasma scale length. Note that for these preplasma parameters, absorption can also be enhanced due to 2D effects
originating from the deformation of the irradiated target surface [30]. This leads us to the conclusion that in our
conguration, the standing wave heating mechanism is the most efcient to generate hot electrons.
2.3. Proton acceleration
We now plot in Fig. 1c) the cut-off proton energies that result from the laserfoil interaction with various density
scale lengths. These energies correspond to saturated values obtained after several picoseconds. First, as expected,
the proton cut-off energy increases as the laser energy increases, as already measured [31]. On the whole, the plot of
proton maximum energy versus the preplasma length is similar to the absorption curve.
Fig. 1c) shows that for a given laser energy, increasing the proton cut-off energy can be achieved by increasing
the plasma scale length. This shows that in the TNSA regime of proton acceleration, a two-pulse conguration (using
a low-intensity beam to control the plasma gradient before sending the high-intensity beam onto the target) is able
to increase the peak proton kinetic energy. This setup involves the generation of a preplasma in front of the thin foil
target in order to enhance the fraction of absorbed laser energy, and a knowledge of the shock wave propagation to
determine the maximum time delay at which the relativistic laser beam should be focused on target.
We have observed that the electron heating mechanisms occurring during the laserplasma interaction depend
on both the laser intensity and the plasma characteristic length. For short plasma length, the electrons are mainly
accelerated along the laser propagation direction by the combined effect of the oscillating Lorentz force and restoring
electrostatic eld due to the plasma ions [24,25]. If the laser intensity is sufciently high and the plasma gradient not
too steep, plasma waves can be resonantly excited in the overdense target by the Lorentz force, leading to absorption
enhancement [26]. Yet, when the preplasma becomes larger, stochastic heating by the standing electromagnetic wave
eventually dominates [29], yielding the maximum absorption rate and the hottest electrons. Although the electrostatic
accelerating eld in this heating regime is slightly nonuniform, the protons reach higher kinetic energy when such a
large preplasma is present in front of the target.
This study suggests that, in the proposed two-pulse setup, the delay between the laser pulses should be just below
the transit time of the shock wave through the target. This would maintain a sharp gradient at the target rear surface
and allow maximum extension (over distances of several m) of the front surface plasma, thereby optimizing electron
heating and proton acceleration.
3. Dependence of acceleration on pulse duration
Alternatively, optimization of ion energy for a xed incident laser energy can be obtained by varying the laser
parameters, and chiey its pulse duration. Indeed, one can imagine that a very short and intense pulse would generate
a relatively small number of very energetic electrons. On the contrary, we could expect a large number of less ener-
getic electrons from a lower-intensity, longer-duration pulse. How the acceleration varies between these two limits is
still a relatively open question. In addition, effects such as variations of the absorption with laser intensity or target
E. Lefebvre et al. / C. R. Physique 10 (2009) 197206 201
Fig. 2. Maximum proton energy and absorption vs pulse duration for the smooth-edge (a) and sharp-edge (b) target proles. (c) Gradient length at
rear surface, at peak of pulse, vs pulse duration.
preheating by the rising edge of the pulse can affect the position of the optimum pulse duration for the maximum
proton energy between these two limits.
We show in this section that varying the pulse duration actually inuences the maximum proton energy through
three mechanisms: the characteristic gradient length at the target rear surface, the characteristic time of energy ex-
change between protons and electrons, and the laser absorption rate. Overall, we observe relatively moderate variations
of the peak proton energy. This weak dependance, however, is found to result from strong but conicting variations of
the above-mentioned parameters.
3.1. Numerical results
In order to study the effect of pulse duration on the maximun proton energy, we carried out another set of 2D
simulations using CALDER. In these simulations, the p-polarized laser pulse with a wavelength of 0.8 m is emitted
from the left side of the simulation box with an incidence angle of 45

on target and a focal spot of 2.8 m. The pulse


temporal and spatial proles are Gaussian. The temporal full width at half maximum,
p
, is varied from 30 fs to 300 fs.
The pulse energy is kept constant and equal to 0.14 J. Consequently, its intensity varies from 3.6 10
19
W/cm
2
to
3.6 10
18
W/cm
2
. The peak intensity enters the box at a time equal to 1.83 times the FWHM, so that the rising edge
of the main pulse can be fully described. In our simulations, we used either a sharp-edge or a smooth-edge target, the
latter with an exponential preplasma at the front side of the form
n
e
(x) =50n
c
exp

x
l
ss

l
ss
is the characteristic gradient length and the preplasma density therefore goes from 1% to 100% of 50n
c
over 4.6l
ss
.
The target is fully-ionized hydrogen. The numerical parameters and boundary conditions used are similar to those of
the previous section.
All simulations have been run to roughly 600 fs after the peak of the laser pulse reaches the target front surface, a
point where we are condent that subsequent evolution of ion peak energy will not qualitatively change our conclu-
sions. For all the pulse durations we considered, we plot in Figs. 2a) and 2b) the maximum proton energy at t =500 fs
and the absorption as a function of pulse duration,
p
. For both target proles, we notice a steady increase of absorp-
tion with pulse duration: from 50% for
p
=30 fs to 70% for
p
=300 fs for the smooth-edge target, and even larger
variations for the sharp-edge target, from 22% for the shortest pulses to 70% for the longest ones. In comparison, the
variations of the maximum proton energy are more subtle. For smooth-edge targets, it rst increases slightly from
11 to 13 MeV between
p
= 30 fs and 84 fs, then decreases down to 8 MeV as the pulse duration becomes longer.
A similar trend is observed for the other target prole, but the peak proton energy now only reaches roughly 10 MeV
at
p
=219 fs before decreasing for larger pulse durations.
202 E. Lefebvre et al. / C. R. Physique 10 (2009) 197206
3.2. Discussions
During the rising edge of the pulse, signicant electron heating starts around 10
17
W/cm
2
and the hot electrons
expand around the target. When the laser pulse duration increases, the front surface plasma heated by the rising edge
of the pulse has more time to expand before the peak intensity hits the target. As the electron density gradient length
increases at the front side, absorption increases [32,19] and more electrons are heated to high temperature. This effect
is more apparent for the sharp-edge targets, for which no preplasma is initially present. These heated electrons travel
through the target, come out at the rear side and expand into the vacuum, creating an electrostatic eld which starts
accelerating protons located at the back surface. Thus, protons begin to be accelerated before the arrival of the peak
of the pulse at the target front surface. This early proton acceleration creates a density gradient expanding from the
back surface into vacuum. This gradient is known to diminish the strength of the electrostatic eld seen by the protons
located at the back surface [22,33]. In Fig. 2c), the back surface gradient length at the time of peak intensity is plotted
as a function of the laser pulse duration for both target proles. It can be seen that the gradient length increases with

p
. This is related to the expansion time of protons which becomes larger when the laser pulse duration increases.
This gradient created at the rear side should hamper proton acceleration [22]. However one can notice fromFigs. 2a)
and 2b) that the maximumproton energy rst increases with
p
before decreasing. We are actually facing a competition
between absorption, screening by the rear side density gradient, and variations in the characteristic energy exchange
time between electrons and protons. While the gain in proton energy due to absorption is higher than the loss due to
electrostatic shielding at the rear surface, the maximum proton energy increases. On the contrary, when the energy
gain due to absorption cannot overcome the loss due to shielding anymore, we observe a decrease in the maximum
proton energy. This balance points out to the existence of an optimum pulse duration. This optimum is found to be
relatively weak, and depends on the details of the interaction: for instance, it varies from 84 fs to 219 fs when the
target prole is changed from smooth-edge to sharp-edge in the above calculations.
3.3. Conclusion
In this study of pulse duration inuence on ion acceleration, as well as more detailed analyses that will be published
elsewhere [18], we identied three processes through which a variation of pulse duration inuences the maximum
proton energy: the preplasma expansion at the front side of the target [17], the early rear side expansion due to
electron heating before the arrival of the pulse peak, and the acceleration time. We have shown that despite strong
variations of each of these parameters, they eventually combine to result in relatively moderate variations of the peak
proton energy as a function of pulse duration.
4. Isochoric heating with protons
To ensure uniform ion energy deposition through a thin target, it is possible to take advantage of the variations of
the proton stopping power with energy. Ions are known to deposit their energy more efciently towards the end of
their range, in the so-called Bragg peak, whereas the stopping power variations at high energy are relatively moderate.
A palatable idea is then to tailor the accelerated ion distribution in order to suppress low-energy protons and make sure
that all protons incident on the second target are energetic enough to be transmitted through it with little variations of
their stopping power along the target thickness. Modication of the proton spectrum has been predicted or observed
to result from a variation of the deposited proton layer at the target rear surface, from a variation of the proton doping
fraction in multi-ion species targets [34], and from a variation of the proton layer radius when 2D geometry effects
are concerned. We will review below the effect of these parameters.
4.1. Proton acceleration
We will rst use the CALDER code to compute proton acceleration in 2D geometry off a thin, dense target, with
various types of proton layers deposited on its back surface. In all these calculations, the incident laser pulse has a
0.8 m wavelength, is linearly polarized and incident from the lower left corner of the simulation box in p polarization,
with an angle of 45

on target. The pulse duration is 65 fs full width at half maximum (FWHM) and the focal
spot diameter is 8 m FWHM. The normalized laser eld amplitude is a
0
= 1.6, corresponding to an intensity of
E. Lefebvre et al. / C. R. Physique 10 (2009) 197206 203
Fig. 3. Laser intensity and ion densities (heavy ions and protons) throughout the simulation at t =93, 245, 361, and 513 fs.
Fig. 4. Proton bunch densities at the end of the simulations (t =513 fs) for three initial proton layer diameters: 13, 51, and 77 m. (The top of the
color scale is at 0.1n
c
.) The initial proton layer position is x =16.4 m.
5.5 10
18
W/cm
2
. This laser pulse is incident onto a 140 nm-thick, 100 n
c
bulk target, with an ion mass-to-charge
ratio equal to ten times that of protons.
In a rst series of simulations, we considered a proton layer of 10 nm thickness at the back surface of this target,
with a density of 10 n
c
, and varied its radial extent between 13, 18, 51, and 77 m the latter value being equal to
the whole transverse dimension of the simulation box. In a second series of simulations, we kept the diameter of the
proton layer equal to 18 m, but varied instead its density, from 5 to 50 n
c
. In a last simulation, the proton layer was
replaced by a homogeneous mixture of heavy ions and protons throughout the whole target, with a 1% fraction of
protons.
The simulation setup and geometry are illustrated in the rst panel of Fig. 3, where the incident laser pulse and
target ion densities are superimposed early in the simulation, at t =93 fs. The incident pulse has barely reached the
solid target and the proton layer cannot be distinguished from the heavy substrate. Later on, in the second panel taken
at t =245 fs, the incident laser pulse has been fully reected or absorbed, and the protons have moved within 2 m
from their initial location. In the last two panels, at 361 and 513 fs, the proton layer becomes fully detached from the
substrate, as all protons become accelerated to nite energy. The layer becomes thicker and less dense as it accelerates,
because the accelerating eld is not uniform longitudinally throughout the layer. It also picks up some curvature, due
to a similar transverse inhomogeneity.
When the proton layer diameter and density are varied, the parameters of the accelerated proton bunch can change
drastically. The shapes of the proton bunch at the end of the simulations for three diameters are plotted in Fig. 4.
Relatively homogeneous acceleration can be given to layers that are transversely limited to a few times the incident
laser spot. For larger dimensions, the transverse variations of the accelerating eld lead to strong inhomogeneities
in the proton bunch, with some particles located more than 20 m from the center of the target that can be barely
accelerated. This difference is even more visible when the proton layer initially extends across the whole target rear
surface. Its outer edges are then clearly unable to detach from the substrate, whereas all the protons located just behind
the laser impact on target are accelerated to more than 1 MeV. Note that the slight topdown asymmetry in Figs. 3
204 E. Lefebvre et al. / C. R. Physique 10 (2009) 197206
Fig. 5. Proton energy distributions at the end of the simulations (t =513 fs) for three initial proton layer densities: 5, 10, and 50 n
c
.
and 4 can be attributed to the imperfect vertical alignment between the center of the proton layer, the center of the
simulation box, and the laser impact on target.
Consistent with these observations, the proton energy distributions for the wide or dense layers show broad spectra
extending up to zero kinetic energy, where a maximum number of particles are located. On the other hand, for small
enough layers (13 and 18 m diameters) as well as for the lower proton layer densities (5 and 10 n
c
), a non-zero
minimum cutoff energy is observed, and the position of this cutoff is found to increase when thinner or less dense
proton layers are used. This trend is clearly apparent in the three panels of Fig. 5 where the distributions are compared
for a 18 m diameter layer at 5, 10, and 50 n
c
density. But one also observes that as the minimum energy cutoff moves
to higher values, the overall number of protons at a given energy is also reduced. In other words, using thinner and
thinner proton layer densities enables to produce proton bunches without the low-energy particles that are detrimental
for uniform energy deposition, but at the same time this acceleration is less and less efcient.
Finally, let us mention that if the last target type, in which the protons are uniformly distributed inside the substrate
initially, does indeed lead to a spectrum showing some modulations as previously reported [34], the overall efciency
and high-energy particle numbers are even lower, and a large amount of low-energy protons can still be observed,
suggesting that this type of targets is not appropriate for heating experiments without due optimization.
4.2. Heating of a thin aluminum layer
The canonical secondary target that we consider below is a 10 m aluminum target, a case already considered
experimentally. The front surface of this target is separated by 246 m of vacuum from the back of the laser-irradiated,
proton-producing target. We use here the CMC particle transport code to compute the propagation and slowing-down
of protons in this solid target, and the energy they deposit along their paths. This code and the stopping powers it
uses have been documented in a previous publication [16]. The self-consistent evolution of stopping powers as a
function of target heating and ionization is not accounted for here: only the cold stopping powers for aluminum are
used [35]. Protons accelerated in the CALDER PIC calculations are input in the CMC calculation with their specic
positions and momenta. Their trajectories are then followed until they hit the target, slow down and deposit part of
their kinetic energy in the aluminum layer. Typical proton trajectories and a map of the resulting deposited energy are
plotted in Fig. 6a), corresponding to the above PIC calculation with a 10 n
c
, 18 m diameter proton layer at the back
surface of the laser-irradiated target. High energy protons arrive rst on target and are energetic enough to be rapidly
transmitted. Therefore, energy deposition at the front and rear target surfaces initially follow similar variations. Later
on, as lower-energy protons reach the secondary target, these particles are not energetic enough to be transmitted
through the target. Therefore, the front and rear surface specic energies are decoupled, with the front surface being
heated for several more picoseconds by low-energy protons. This trend is apparent in Fig. 6b), where the variations
of specic energies with time are plotted for the peak value across the target, and two points close to the laser axis on
the front and rear surfaces.
With little surprise, we observe that similar coupled calculations for lower-density proton layers, as considered in
Fig. 5, lead to less decoupling between the target front and rear surfaces, as less low-energy protons are present in
E. Lefebvre et al. / C. R. Physique 10 (2009) 197206 205
Fig. 6. (a) Trajectories of protons emitted by the laser-irradiated target (not shown), that propagate from its back surface and heat the secondary
aluminum target; the color code indicates the specic energy deposited in the target; (b) Specic energy deposited in the aluminum layer by the
incident proton beam. Peak value (red) and values on two points at the front and back target surfaces.
the incident beam. For instance, the ratio of front to back surface specic energies is E
front
/E
back
2.5 in Fig. 6b),
obtained for a 10 n
c
, 18 m diameter proton layer at the rst target back surface. This ratio increases to E
front
/
E
back
> 4 for a 20 n
c
layer, and decreases to roughly 2 for a 5 n
c
layer. However, as the mismatch between front and
rear surfaces is reduced with lower layer densities, so is also the absolute value of specic energy, as less protons are
incident on the secondary target.
5. Conclusions
Proton acceleration from the rear surface of thin laser-irradiated solid foils depends on a number of parameters
such as target density prole and thickness, laser intensity, energy, pulse duration, polarization and angle of incidence.
The intricacies of these dependences only begin to be elucidated. Two strategies for optimizing the peak proton energy
were followed in this article. We rst showed that under certain conditions, a controlled low-intensity prepulse could
be helpful by preforming a smooth density prole at the irradiated surface of the target and hence increasing the
high-intensity laser absorption. Our calculations suggest that the maximum delay between the low- and high-intensity
pulses should be used that is compatible with an unperturbed target back surface. Alternatively, lengthening the high-
intensity pulse duration, even if it comes at the expense of its peak intensity, may also increase the peak proton energy
as the early part of the pulse heats up the front target surface and increases the absorption of the trailing part. Some
optimum pulse duration can be found for which this effect dominates the reduction of proton energy due to screening
of the back surface electrostatic eld by the density gradient that also develops there.
Some preliminary calculations of heating by these laser-produced proton beams were nally reported. They showed
that a low-energy cutoff can be created in the proton distribution by reducing the proton layer radius and density, albeit
at the expense of the overall acceleration efciency. Proton layers that are either too thick or larger than a few times
the laser spot radius lead to proton distributions that show no low-energy cutoff and therefore result in very non-
uniform energy deposition in the secondary target. On the other hand, using properly optimized proton-producing
targets irradiated with typical 30 fs laser pulses at a few Joule level, one should be able to heat aluminum to GJ/kg
specic energy with interesting uniformity across 5 m-thick samples.
206 E. Lefebvre et al. / C. R. Physique 10 (2009) 197206
References
[1] V. Yanovsky, V. Chvykov, G. Kalinchenko, P. Rousseau, T. Planchon, T. Matsuoka, A. Maksimchuk, J. Nees, G. Chriaux, G. Mourou, K.
Krushelnick, Opt. Express 16 (2008) 2109.
[2] T.E. Cowan, J. Fuchs, H. Ruhl, A. Kemp, P. Audebert, M. Roth, R. Stephens, I. Barton, A. Blazevic, E. Brambrink, J. Cobble, J. Fernndez,
J.-C. Gauthier, M. Geissel, M. Hegelich, J. Kaae, S. Karsch, G.P. Le Sage, S. Letzring, M. Manclossi, S. Meyroneinc, A. Newkirk, H. Ppin,
N. Renard-LeGalloudec, Phys. Rev. Lett. 92 (2004) 204801.
[3] J. Fuchs, T.E. Cowan, P. Audebert, H. Ruhl, L. Gremillet, A. Kemp, M. Allen, A. Blazevic, J.-C. Gauthier, M. Geissel, M. Hegelich, S. Karsch,
P. Parks, M. Roth, Y. Sentoku, R. Stephens, E.M. Campbell, Phys. Rev. Lett. 91 (2003) 255002.
[4] M. Borghesi, D.H. Campbell, A. Schiavi, M.G. Haines, O. Willi, A.J. MacKinnon, P. Patel, L.A. Gizzi, M. Galimberti, R.J. Clarke, F. Pegoraro,
H. Ruhl, S. Bulanov, Phys. Plasmas 9 (2002) 2214.
[5] P.K. Patel, A.J. Mackinnon, M.H. Key, T.E. Cowan, M.E. Foord, M. Allen, D.F. Price, H. Ruhl, P. Springer, R. Stephens, Phys. Rev. Lett. 91
(2003) 125004.
[6] S.V. Bulanov, V.S. Khoroshkov, Plasma Phys. Rep. 28 (2002) 453.
[7] M. Roth, T.E. Cowan, M.H. Key, S.P. Hatchett, C. Brown, W. Fountain, J. Johnson, D.M. Pennington, R.A. Snavely, S.C. Wilks, K. Yasuike,
H. Ruhl, F. Pegoraro, S.V. Bulanov, E.M. Campbell, M.D. Perry, H. Powell, Phys. Rev. Lett. 86 (2001) 436.
[8] B.M. Hegelich, B.J. Albright, J. Cobble, K. Flippo, S. Letzring, M. Paffett, H. Ruhl, J. Schreiber, R.K. Schulze, J.C. Fernndez, Nature 439
(2006) 441.
[9] H. Schwoerer, S. Pfotenhauer, O. Jckel, K.-U. Amthor, B. Liesfeld, W. Ziegler, R. Sauerbrey, K.W.D. Ledingham, T. Esirkepov, Nature 439
(2006) 445.
[10] T. Toncian, M. Borghesi, J. Fuchs, E. dHumires, P. Antici, P. Audebert, E. Brambrink, C.A. Cecchetti, A. Pipahl, L. Romagnani, O. Willi,
Science 312 (2006) 410.
[11] A.P.L. Robinson, D. Neely, P. McKenna, R.G. Evans, Plasma Phys. Controlled Fusion 49 (2007) 373.
[12] N. Kumar, A. Pukhov, Phys. Plasmas 15 (2008) 053103.
[13] P. Antici, J. Fuchs, S. Atzeni, A. Benuzzi, E. Brambrink, M. Esposito, M. Koenig, A. Ravasio, J. Schreiber, A. Schiavi, P. Audebert, J. Phys.
IV France 133 (2006) 1077.
[14] R.A. Snavely, B. Zhang, K. Akli, Z. Chen, R.R. Freeman, P. Gu, S.P. Hatchett, D. Hey, J. Hill, M.H. Key, Y. Izawa, J. King, Y. Kitagawa,
R. Kodama, A.B. Langdon, B.F. Lasinski, A. Lei, A.J. MacKinnon, P. Patel, R. Stephens, M. Tampo, K.A. Tanaka, R. Town, Y. Toyama, T.
Tsutsumi, S.C. Wilks, T. Yabuuchi, J. Zheng, Phys. Plasmas 14 (2007) 092703.
[15] M.E. Foord, P.K. Patel, A.J. Mackinnon, S.P. Hatchett, M.H. Key, B. Lasinski, R.P.J. Town, M. Tabak, S.C. Wilks, High Energy Density
Phys. 3 (2007) 365.
[16] E. Lefebvre, E. dHumires, S. Fritzler, V. Malka, J. Appl. Phys. 100 (2006) 113308.
[17] R. Nuter, L. Gremillet, P. Combis, M. Drouin, E. Lefebvre, A. Flacco, V. Malka, J. Appl. Phys. 104 (2008) 103307.
[18] M. Carri, E. Lefebvre, A. Flacco, V. Malka, in preparation.
[19] E. Lefebvre, G. Bonnaud, Phys. Rev. E 55 (1997) 1011.
[20] E. Lefebvre, N. Cochet, S. Fritzler, V. Malka, M.-M. Alonard, J.-F. Chemin, S. Darbon, L. Disdier, J. Faure, A. Fedotoff, O. Landoas, G.
Malka, V. Mot, P. Morel, M. Rabec Le Gloahec, A. Rouyer, Ch. Rubbelynck, V. Tikhonchuk, R. Wrobel, P. Audebert, C. Rousseaux, Nucl.
Fusion 43 (2003) 629.
[21] J.-P. Colombier, P. Combis, A. Rosenfeld, I.V. Hertel, E. Audouard, R. Stoian, Phys. Rev. B 74 (2006) 224106.
[22] A.J. Mackinnon, M. Borghesi, S. Hatchett, M.H. Key, P.K. Patel, H. Campbell, A. Schiavi, R. Snavely, S.C. Wilks, O. Willi, Phys. Rev.
Lett. 86 (2001) 1769.
[23] T.-Y. Brian Yang, W.L. Kruer, R.M. More, A.B. Langdon, Phys. Plasmas 2 (1995) 3146.
[24] P. Gibbon, E. Frster, Plasma Phys. Controlled Fusion 38 (1996) 769.
[25] P. Mulser, D. Bauer, H. Ruhl, Phys. Rev. Lett. 101 (2008) 225002.
[26] A.A. Andreev, K.Yu. Platonov, T. Okada, S. Toraya, Phys. Plasmas 10 (2003) 220.
[27] Z.-M. Sheng, K. Mima, Y. Sentoku, M.S. Jovanovi c, T. Taguchi, J. Zhang, J. Meyer-ter-Vehn, Phys. Rev. Lett. 88 (2002) 055004.
[28] A. Bourdier, D. Patin, E. Lefebvre, Physica D 206 (2005) 1.
[29] Y. Sentoku, V.Y. Bychenkov, K. Flippo, A. Maksimchuk, G. Mourou, Z.M. Sheng, D. Umstadter, Appl. Phys. B 74 (2002) 207.
[30] H. Ruhl, A. Macchi, P. Mulser, F. Cornolti, S. Hain, Phys. Rev. Lett. 82 (1999) 2095.
[31] Y. Oishi, T. Nayuki, T. Fujii, Y. Takizawa, X. Wang, T. Yamazaki, K. Nemoto, T. Kayoiji, T. Sekiya, K. Horioka, Y. Okano, Y. Hironaka, K.G.
Nakamura, K. Kondo, A.A. Andreev, Phys. Plasmas 12 (2005) 073102.
[32] S. Wilks, W. Kruer, M. Tabak, A.B. Langdon, Phys. Rev. Lett. 69 (1992) 1383.
[33] T. Grismayer, P. Mora, Phys. Plasmas 13 (2006) 032103.
[34] A.P.L. Robinson, A.R. Bell, R.J. Kingham, Phys. Rev. Lett. 96 (2006) 035005.
[35] J.F. Ziegler, J. Appl. Phys. 85 (1999) 1249.

You might also like